Extension of the Vorcat Technology to Moving Boundaries

Turbulence associated with moving boundaries is notoriously difficult to model and no more so than using traditional Reynolds Averaged Navier Stokes (RANS) closures that have at best a tenuous connection to the underlying physics. While many industrial users of ComputationalFluid Dynamics (CFD) employ codes based on solving RANS equations in the treatment of flows with moving boundaries this is mostly due to the absence of credible alternatives. There is clear interest in industry for the development of rapid simulation techniques that use better physics to achieve better predictions.

An ongoing widespread international effort aims to perfect schemes for computing turbulent flow on supercomputers by large eddy simulation (LES) methods in which the large scales of motion are simulated and the effect of the smallest scales on the large, (i.e., those that fall beneath the size of the numerical mesh) are modeled. In LES the transient details of the flow are computed, so the method is particularly well suited for complex problems including those with moving boundaries where vortex shedding and other phenomena need to be accounted for. Though some LES methods have been incorporated in commercial CFD codes, such schemes rely on overly diffusive models that seriously limit their effectiveness; considerable interest remains in finding better models.

The Vorcat, Inc. turbulent flow simulation software is an unorthodox, LES approach incorporating a hybrid vortex filament/finite volume scheme that is designed to circumvent many of the drawbacks of traditional methods depending on a numerical mesh. Vorcat is readily executed on HPCs and may be potentially adapted to small clusters equipped with GPUs. In Vorcat, freely convecting and interacting vortex filaments are used away from boundaries to directly account for the rotational motions at many scales that are fundamental to turbulent flow.This preserves the physics of vortices in the flow that would otherwise by compromised by diffusion in grid-based schemes. Next to boundaries Vorcat provides for high resolution calculation of vorticity generation on a prismatic mesh where the concept of a LES is inappropriate.

A number of validation studies have been successful in establishing the unique benefits and accuracy of the Vorcat approach. Of particular significance has been the simplicity with which complex flows can be treated in their natural setting including spatially growing boundary layers and shear layers. Besides quantitative accuracy, the solutions have provided important new insights into the structure of turbulent flow. Once the additional capability of accommodating moving boundaries is developed and implemented within Vorcat, the approach will find immediate application in the design of many new energy technologies.